Again, the proportions of railway traffic connected with foreign trade may vary greatly, and it may happen that a big increase in railway gross receipts is due to increasing foreign trade, primarily. There is reason to suppose that much of the increase of 1916 is to be explained that way. This makes our comparison for 1916 particularly adverse to foreign trade, since we count as domestic trade what is really foreign trade. The figures, however, are presented as they stand. Moreover, for 1916, the great increase in foreign trade is in _exports_. Merchandise imports are not much greater than in previous years.[309] Our exports have been chiefly paid for by "invisible items," gold and securities, and short term credits. These do not appear anywhere in our figures. A substantial source of error appears from this cause in our 1916 figure. I should think it safe to put the ratio for foreign trade to domestic trade for 1916 at above 20%, instead of the 17.9% our table shows. The reader will wish to know for a given year how much of the increase or decrease is due to physical growth of business, as represented by railway gross receipts, and how much is due to changes in prices. To give this information, and to make it easy for a critic to check the results, a table showing the index numbers from which the figures for net income are computed is subjoined.[310] TABLE I[311] 1 2 3 4 Ratio of Domestic Trade of Foreign Trade of Foreign Calendar Net Income United States = United States = to Years of the Net Income minus Exports at Retail Domestic United Imports at Retail Prices Trade States Prices